Predictive IT Ticketing via Machine Learning Feature Extraction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing IT solutions for addressing IT issues rely heavily on user-provided information, which can be inaccurate and lead to misclassification of issues, delayed problem recognition, and inefficient manual processes, resulting in prolonged issue resolution times and inadequate urgency assessment.

Innovation Solution

Implementing a predictive ticketing system that uses machine learning to extract features from monitoring data, apply a machine learning model to identify suitable insights, and generate predictive tickets with textual descriptions of expected future symptoms, allowing for preemptive issue detection and resolution without user input.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Extent of automation

If user-provided information is used for IT issue classification, then the system can operate with simple automated processes, but the accuracy of issue classification deteriorates due to inaccurate user descriptions

Engineering Contradiction:

Improveautomated ticket processingVSAvoidissue classification accuracy

Core Design Contradiction:

Extent of automationVSMeasurement precision

Solution Approach 1:

The patent introduces machine learning models as an intermediary between user-reported symptoms and IT issue classification. The ML models process monitoring data and user inputs to generate accurate issue classifications, eliminating the direct dependency on potentially inaccurate user descriptions while maintaining automated operation.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces manual classification processes with machine learning-based automated classification. The ML models analyze monitoring data and user inputs to automatically classify issues, substituting human operators and eliminating the accuracy problems associated with manual classification while maintaining high automation levels.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If manual playbooks are used for IT issue resolution, then the system can handle complex branching scenarios, but the resolution time increases due to manual investigation steps

Engineering Contradiction:

Improvehandling complex issue scenariosVSAvoidissue resolution time

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-processing monitoring data and identifying potential issues before they manifest as user-reported problems. The system continuously analyzes monitoring data and generates predictive tickets, allowing IT teams to address issues proactively rather than reactively, significantly reducing resolution times while maintaining the ability to handle complex scenarios.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback mechanisms where the system continuously monitors IT infrastructure, compares current state against historical data and thresholds, and automatically generates tickets when anomalies are detected. This closed-loop feedback system enables rapid response to issues while maintaining adaptability to handle complex branching scenarios through learned patterns.

Inventive Principle:

Principle #23Feedback

3Use of energy by moving object

If reactive ticketing based on user reports is used, then the system requires minimal monitoring resources, but issue detection is delayed until users report problems

Engineering Contradiction:

Improvemonitoring resource consumptionVSAvoidissue detection time

Core Design Contradiction:

Use of energy by moving objectVSLoss of time

Solution Approach 1:

The patent applies partial action by implementing selective monitoring and analysis only when anomalies are detected or when predictive models identify potential issues. Rather than continuously analyzing all monitoring data at full capacity, the system uses ML models to filter and prioritize events, maintaining low resource consumption while enabling proactive issue detection that reduces detection time.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS11416325B2Machine-learning and deep-learning techniques for predictive ticketing in information technology systems

Publication Date: 2022.08.16 SERVICENOW INC

AI summary

A system and method for predictive ticketing in information technology (IT) systems. The method includes extracting a plurality of features from monitoring data related to an IT system, wherein the plurality of features includes at least one incident parameter, wherein the monitoring data includes machine-generated textual data; applying a machine learning model to the extracted plurality of features, wherein the machine learning model is configured to output a suitable insight for an incident represented by the at least one incident parameter, wherein the suitable insight is selected from among a plurality of historical insights; and generating a predictive ticket based on the suitable insight, wherein the predictive ticket includes a textual description of an expected future symptom in the IT system.